Forecasting Misused E-Commerce Consumer Returns

High return rates are a problem particularly concerning e-tailers. Often, these returns are considered to come back as good as new which enables an immediate resale. This assumption does not take into account that some consumers show a dishonest buying intention by returning items they have already used. For reasons of an improved operational planning it could be helpful to predict this kind of product returns. Therefore, we adapt and enhance already existing forecasting approaches to situations where the e-tailer matches consumer returns with their date of sale and furthermore, where the consumer has to register the return first. The so generated product-individual data are covered by developing three forecasting methods. It is examined under which circumstances the models show a satisfying accuracy and correspondent requirements that have to be met in this context are inferred. The overall premise is the practicability in order to ensure an uncomplicated application.